Are evaluations just for Text-text, or is there an efficient ways to evaluate image-text, like for MobileClip2, or YOLOE? At first blush, it seems like that would just be redoing training, testing, and validation of the dataset, or am I missing something? (newbe)

Is Evaluations only for swift or does it support other languages like python and others?

Performance wise what are trade-offs when running an MLX-backed model on-device compared to using the system's AFM Core model? Also semiconnected: How do I use the 'model judge evaluator' to compare the accuracy of a custom LoRA adapter against the system's private cloud compute models?